Network analysis of promoter interactions reveals the hierarchical differences in genome organisation between human pluripotent states

Abstract

SUMMARYA complex and poorly understood interplay between 3D genome organisation, transcription factors and chromatin state underpins cell identity. To gain a systems-level understanding of this interplay, we generated a high-resolution atlas of annotated chromatin interactions in naïve and primed human pluripotent stem cells and developed a network-graph approach to examine the atlas at multiple spatial scales. Investigating chromatin interactions as a network uncovered highly connected hubs that changed substantially in interaction frequency and in transcriptional co-regulation between pluripotent states.Small hubs frequently merged to form larger networks in primed cells, often linked by newly-formed Polycomb-associated interactions. Importantly, we identified state-specific differences in enhancer activity and interactivity that corresponded with widespread reconfiguration of transcription factor binding and target gene expression. These findings provide multilayered insights into the gene regulatory control of human pluripotency and our systems-based network approach could be applied broadly to uncover new principles of 3D genome organisation. RESULTS Promoter interaction mapping in naïve and primed human PSCsWe used PCHi-C to profile the global, high-resolution interactomes of 22,101 promoters in isogenic human naïve and primed PSCs. There was a strong concordance in pairwise interaction read counts between the biological replicates of the same cell type (r 2 >0.95; Figure S1A); therefore we merged replicates for all downstream analyses. PCHi-C data normalisation and signal detection using the CHiCAGO pipeline (Cairns et al., 2016) identified 75,091 significant cis-interactions between baited promoters and other genomic regions in naïve PSCs, and 83,782 in primed PSCs ( Figure S1B). Just under half of the interactions were common to both cell types (n=39,360). As expected, trans-interactions represented a small minority of promoter interactions (354 interactions). In both cell types, the majority of significant interactions were between the promoters of protein-coding genes and non-promoter genomic regions ( Figure S1B).Processed datasets from this large-scale resource are available through the Open Science Framework (https://osf.io/XXXXX) and Table S1, and raw sequencing reads have been deposited to the Gene Expression Omnibus (accession GSEXXXXXX). Network visualisation of promoter interactomesWe next developed a computational approach called Canvas (Chromosome architecture network visualisation at scales) to visualise high-resolution, capture-based DNA interaction data at a network scale.Network graphs were constructed where each node of the network represents an individual HindIII genomic fragment (average size, 4 kb) and each edge represents a significant interaction between nodes ( Figure 1A). We combined all significant interactions detected in naïve and primed PSCs to produce a single, unified network graph, which retains information about whether an interaction is shared or cell type-specific ( Figu...

Chromatin Profiling of Human Naïve Pluripotent Stem Cells

“…Early ChIP-sequencing experiments investigated a range of transcription factors, such as STAT1 [1,2], histone modifications, and transcriptional apparatus proteins, such as RNA polymerase II [3]. More recently, ChIP-sequencing has been performed in naı ¨ve hPSCs for the core pluripotency transcription factors NANOG, SOX2, and OCT4 [4,5] among a number of other transcription factors [5][6][7][8] as well as histone post-translational modifications, including H3K4me1, H3K4me3, H3K9me3, H3K27me3, and H3K27ac [3,4,6,7,9,10]. These studies have provided extensive information about gene regulation by the pluripotency network and have identified naı ¨ve-specific gene regulatory elements, such as enhancers and promoters and their epigenetic states.…”